Solid State disk for a desktop system C drive?

A

Al Dykes

Is there any reason why I shouldn't use a sold state disk as a C drive
on a high performance desktop system? 36GB would make a fine C drive
for my work.

Newegg has 36GB solid state disks starting at 70 bucks. They feature
screaming read rates. The write rates are a little slower but still
faster than a 7200RPM hard disk.

Latency should be zero but I know that something has to happen in the
write cycle for a SSD and the summary specs don't list delay times
which would be a form of latency.


I do Photoshop. I'm thinking of using the SSD for a second disk and
dedicating it to swap, temp, and PS work files. That way, if it dies,
I haven't lost any work.

comments?
 
A

Arno Wagner

Previously Al Dykes said:
Is there any reason why I shouldn't use a sold state disk as a C drive
on a high performance desktop system? 36GB would make a fine C drive
for my work.

I don't see any. In fact I have done this recently witn an OCZ
30GB drive. Not that much faster, but noticeable.
Newegg has 36GB solid state disks starting at 70 bucks. They feature
screaming read rates. The write rates are a little slower but still
faster than a 7200RPM hard disk.
Latency should be zero but I know that something has to happen in the
write cycle for a SSD and the summary specs don't list delay times
which would be a form of latency.

The one thing slower is small accessed. I believe these SSDs
have something like 128kB sectors (internally, not visible).
So if you do very small writes or reads, the data rate may drop
significantly.
I do Photoshop. I'm thinking of using the SSD for a second disk and
dedicating it to swap, temp, and PS work files. That way, if it dies,
I haven't lost any work.

Also a good idea. However SSDs should be at least one order of
magnitude more reliable that well-treaded HDDs.
comments?

Go for it.

Arno
 
A

Arno Wagner

Have you tried google before posting?

I think this is now mostly obsolete information. Unless you
have a system that swaps a lot, you are not going to hit the
FLASH life expectancy within 5 years or so. Incidentially
you can get SSDs with 5 year and more warranty time.

Some ElCheapo SSDs may still be affected though. The solution
should be to get a better SSD IMO.

Arno
 
R

Rod Speed

Al Dykes wrote
Is there any reason why I shouldn't use a sold state disk
as a C drive on a high performance desktop system?

Depends on the technology and the OS.

Win and Vista both use it pretty intensively for temporary files and
some solid state disk technologys dont have that may write cycles.
36GB would make a fine C drive for my work.
Newegg has 36GB solid state disks starting at 70 bucks.
They feature screaming read rates. The write rates are
a little slower but still faster than a 7200RPM hard disk.
Latency should be zero but I know that something has
to happen in the write cycle for a SSD and the summary
specs don't list delay times which would be a form of latency.
I do Photoshop. I'm thinking of using the SSD for a second
disk and dedicating it to swap, temp, and PS work files.
That way, if it dies, I haven't lost any work.
comments?

It may not last very long given the limitations on the writes.
 
C

Chris

Is there any reason why I shouldn't use a sold state disk as a C drive
on a high performance desktop system? 36GB would make a fine C drive
for my work.

No reason why you shouldn't... I have an SSD as main drive on my desktop
and works well.
Newegg has 36GB solid state disks starting at 70 bucks. They feature
screaming read rates. The write rates are a little slower but still
faster than a 7200RPM hard disk.

Latency should be zero but I know that something has to happen in the
write cycle for a SSD and the summary specs don't list delay times
which would be a form of latency.

I do Photoshop. I'm thinking of using the SSD for a second disk and
dedicating it to swap, temp, and PS work files. That way, if it dies,
I haven't lost any work.

comments?
I would put the OS on the SSD but it's worth experimenting, as there is
little usage history of these systems. If you do, make sure your backups
are reliable, as they should be.

You may have different priorities, but I have only the SSD drive in that
system and enjoy the silence and the low heat inside the box.

Anyway, have fun and if you don't mind share your experience when you have
time.
 
F

Franc Zabkar

You may have different priorities, but I have only the SSD drive in that
system and enjoy the silence and the low heat inside the box.

Anyway, have fun and if you don't mind share your experience when you have
time.

Speaking of sharing, I'm curious whether SMART is implemented in these
drives, and how? Are there any SMART extensions that are peculiar to
SSDs?

- Franc Zabkar
 
A

Arno Wagner

Speaking of sharing, I'm curious whether SMART is implemented in these
drives, and how? Are there any SMART extensions that are peculiar to
SSDs?

My OCZ does only give temperature and vendor stuff. Goven that
most HDD measures (including reallocated sectors) do not make sense,
this is not surprising. I am a bit disappointed though that it does
not support self-test. I am back to a complete, timed read with error
detection for the SSD.

Arno
 
F

Franc Zabkar

My OCZ does only give temperature and vendor stuff. Goven that
most HDD measures (including reallocated sectors) do not make sense,

How would bad areas of memory be accounted for?

- Franc Zabkar
 
A

Arno Wagner

Previously Franc Zabkar said:
On 19 Jan 2009 13:58:53 GMT, Arno Wagner <[email protected]> put finger
to keyboard and composed:
How would bad areas of memory be accounted for?

The problem here is that SSDs do not use the 512 byte sector
size (instead something much, much larger, like 128kB) and
do not use the HD reallocation mechanism. I would like to
see something like "failed reads" or "ECC recoverd bytes", and
it is possible that the drive actually states them, but in vendor
specific attributes that at least smartctl cannot interpret.
If anybody knows more, below is a slightly shortened SMART
dump of my 30GB OCZ SSD.

Arno


=== START OF INFORMATION SECTION ===
Device Model: OCZ CORE_SSD
Firmware Version: 02.10104
User Capacity: 32,044,482,560 bytes
Device is: Not in smartctl database [for details use: -P showall]
ATA Version is: 7
ATA Standard is: ATA/ATAPI-7 T13 1532D revision 4a
Local Time is: Mon Jan 19 21:59:15 2009 CET
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

SMART Attributes Data Structure revision number: 1280
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 241
9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 0
194 Temperature_Celsius 0x0007 032 100 000 Pre-fail Always - 0
229 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 259986015180268
232 Unknown_Attribute 0x0002 100 048 000 Old_age Always - 9028846498104
233 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 0
234 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 592722311424
235 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 187423
 
F

Franc Zabkar

Previously Franc Zabkar said:
On 19 Jan 2009 13:58:53 GMT, Arno Wagner <[email protected]> put finger
to keyboard and composed:
How would bad areas of memory be accounted for?

The problem here is that SSDs do not use the 512 byte sector
size (instead something much, much larger, like 128kB) and
do not use the HD reallocation mechanism. I would like to
see something like "failed reads" or "ECC recoverd bytes", and
it is possible that the drive actually states them, but in vendor
specific attributes that at least smartctl cannot interpret.
If anybody knows more, below is a slightly shortened SMART
dump of my 30GB OCZ SSD.

Arno


=== START OF INFORMATION SECTION ===
Device Model: OCZ CORE_SSD
Firmware Version: 02.10104
User Capacity: 32,044,482,560 bytes
Device is: Not in smartctl database [for details use: -P showall]
ATA Version is: 7
ATA Standard is: ATA/ATAPI-7 T13 1532D revision 4a
Local Time is: Mon Jan 19 21:59:15 2009 CET
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

SMART Attributes Data Structure revision number: 1280
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 241
9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 0
194 Temperature_Celsius 0x0007 032 100 000 Pre-fail Always - 0
229 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 259986015180268
232 Unknown_Attribute 0x0002 100 048 000 Old_age Always - 9028846498104
233 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 0
234 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 592722311424
235 Unknown_Attribute 0x0002 100 000 000 Old_age Always In_the_past 187423

I would think that SSDs would use wear levelling techniques and that
important indicators of old age would be the number of times that each
memory block had been rewritten. Perhaps one of the attributes
reflects this.

I'd monitor the SMART data on a daily basis and see whether the
attributes change by incrementing the whole word or particular bytes.
Attribute 229 uses all 48 bits, so it appears that it reflects
multiple parameters. OTOH, attribute 235 probably reflects only one
parameter.

- Franc Zabkar
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top